Short intro to the workshop
20 March 2024

Wann bereitstellen, wann archivieren?
Kosten:
Nutzen:
Please check:
rix installed
Nix installed (if not, we’ll do an alternative using github actions)
Alternative to installation of Nix
Do you have git installed?
Do you have a github account?
Compuational Reproducibility: “In principle, all reported evidence should be reproducible. If someone applies the same analysis to the same data, the same result should occur.” (Nosek et al., 2022, p. 721)
Investigated all articles from one issue in Psychological Science: 1 exactly reproducible, 3 essentially reproducible after minor deviations
“Code-specific issues include (a) a lack of shared analysis code or modeling code and (b) issues with package or software versions (often resolvable but sometimes only with considerable effort).”
(Hardwicke et al., 2021) Investigated all articles with open data badges from Psychological Science in 2014-2015:
all target values in 9 out of the 25 articles (36%, CI [19,57]) were reproducible, with the remaining 16 articles (64%, CI [43,81]) containing at least one major numerical discrepancy. After requesting input from original authors, several issues were resolved through the provision of additional information and ultimately all target values in 15 (60%, CI [39,78]) articles were reproducible, with the remaining 10 (40%, CI [22,61]) articles containing at least one major numerical discrepancy. (p. 4)
(Obels et al., 2020) analyzed 62 registered reports. 36 shared both, code and data.
computationally reproduced (21 out of 36, or 58%)
https://doi.org/10.1177/2378023119849803 https://doi.org/10.15626/MP.2018.892

Jürgen Schneider
ju.schneider@dipf.de
Title page Thomas William on Unsplash
Icons by Font Awesome CC BY 4.0

Slides: bit.ly/shared-slides